CN111768413B - Plant three-dimensional point cloud segmentation method and system - Google Patents
Plant three-dimensional point cloud segmentation method and system Download PDFInfo
- Publication number
- CN111768413B CN111768413B CN202010475872.9A CN202010475872A CN111768413B CN 111768413 B CN111768413 B CN 111768413B CN 202010475872 A CN202010475872 A CN 202010475872A CN 111768413 B CN111768413 B CN 111768413B
- Authority
- CN
- China
- Prior art keywords
- point
- stem
- organ
- segmented
- current
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 59
- 230000011218 segmentation Effects 0.000 title claims abstract description 59
- 210000000056 organ Anatomy 0.000 claims abstract description 261
- 239000013598 vector Substances 0.000 claims description 59
- 238000004364 calculation method Methods 0.000 claims description 15
- 238000006243 chemical reaction Methods 0.000 claims description 5
- 238000012847 principal component analysis method Methods 0.000 claims description 5
- 239000000203 mixture Substances 0.000 claims description 3
- 241000196324 Embryophyta Species 0.000 description 64
- 241001057636 Dracaena deremensis Species 0.000 description 19
- 240000008042 Zea mays Species 0.000 description 17
- 235000002017 Zea mays subsp mays Nutrition 0.000 description 13
- 235000005824 Zea mays ssp. parviglumis Nutrition 0.000 description 12
- 235000005822 corn Nutrition 0.000 description 12
- 238000004422 calculation algorithm Methods 0.000 description 5
- 238000010586 diagram Methods 0.000 description 5
- 238000004891 communication Methods 0.000 description 4
- 238000004590 computer program Methods 0.000 description 3
- 210000005069 ears Anatomy 0.000 description 3
- 238000013135 deep learning Methods 0.000 description 2
- 238000001514 detection method Methods 0.000 description 2
- 238000000605 extraction Methods 0.000 description 2
- 238000002372 labelling Methods 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 241000037488 Coccoloba pubescens Species 0.000 description 1
- 235000016383 Zea mays subsp huehuetenangensis Nutrition 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 238000012804 iterative process Methods 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 235000009973 maize Nutrition 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000000877 morphologic effect Effects 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/11—Region-based segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10028—Range image; Depth image; 3D point clouds
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30168—Image quality inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30181—Earth observation
- G06T2207/30188—Vegetation; Agriculture
Landscapes
- Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Apparatus For Radiation Diagnosis (AREA)
- Image Analysis (AREA)
Abstract
The embodiment of the invention provides a plant three-dimensional point cloud segmentation method and system, wherein the method comprises the following steps: dividing a stem point cloud in a three-dimensional point cloud of a target plant; converting the three-dimensional point cloud of the target plant from a global coordinate system to a local coordinate system; extracting the highest point of each non-stem organ in the target plant, and respectively adding the highest point of each non-stem organ into the corresponding initial non-stem organ point set; selecting a target set corresponding to each point to be segmented from the initial stem point set and each initial non-stem organ point set, and adding each point to be segmented into the corresponding target set; and acquiring stems and each non-stem organ in the target plant according to the optimal stem set and each optimal non-stem set. The embodiment of the invention can realize three-dimensional point cloud segmentation of stalks, leaves, tassel and female spike organs in plant plants; and for organ-dense scenes, high-precision segmentation is also possible.
Description
Technical Field
The invention relates to the technical field of computers, in particular to a method and a system for three-dimensional point cloud segmentation of plants.
Background
By utilizing technical means such as a three-dimensional scanner, multi-view three-dimensional reconstruction and the like, three-dimensional point cloud data of corn plants can be obtained, however, how to obtain corn plant phenotype parameters facing agricultural science research through the three-dimensional point cloud data of the plants or how to obtain a three-dimensional grid model facing visual application through the three-dimensional point cloud data of the plants is an important problem in three-dimensional data processing and application of the plants. The method for accurately and efficiently dividing the three-dimensional point cloud of the corn plant to obtain the three-dimensional point cloud data set of the organ scale is an important link in the processing of the three-dimensional point cloud data of the corn plant.
The three-dimensional point cloud segmentation of corn plants mainly comprises the following methods:
(1) Skeleton-based segmentation methods. The method comprises the steps of firstly extracting the skeleton of a corn plant, then dividing the organ skeleton according to skeleton topology, and finally dividing the organ point cloud by utilizing the relation between skeleton points and original point clouds. The method has the difficulty that how to obtain a framework with better quality, the prior art has good effect on plant point clouds with fully unfolded leaves and larger leaf spacing, but under the conditions of more leaves and dense leaves, the framework for segmentation is difficult to extract, and meanwhile, the framework method can not extract corn ears.
(2) A machine learning based method. The method learns the characteristics of the organ point cloud through a large number of manual organ point cloud labels, and classifies and segments the point cloud through a constructed model. The method can only identify corn plant organs similar to the sample, the variety of corn is very large, the morphology difference of the organs is very large, a universal model is difficult to construct for point cloud segmentation, and meanwhile, the current segmentation accuracy of the method is low.
(3) A region growing based method. The method is mainly used for segmenting the stems and the leaves, the regional growth algorithm is used for segmenting the stem point clouds, and then each leaf point cloud is segmented through regional growth from the intersection of the leaves and the stems. Such methods have difficulty in segmenting dense blades by setting parameters.
The three corn point cloud segmentation methods only solve the point cloud segmentation problem of corn stem-leaf organs with large leaf spacing, and when a plurality of leaves are close together and the interval between the leaves is smaller, the high-quality segmentation cannot be performed, and meanwhile, the segmentation of tassel and female spike cannot be performed. Overall, when the organs are very dense, the segmentation error is very large.
Disclosure of Invention
In order to solve the problems, the embodiment of the invention provides a plant three-dimensional point cloud segmentation method and system.
In a first aspect, an embodiment of the present invention provides a method for dividing a three-dimensional point cloud of a plant, including:
dividing a stem point cloud in a three-dimensional point cloud of a target plant, wherein the three-dimensional point cloud of the target plant comprises a stem point cloud and a non-stem organ point cloud, and the divided stem point clouds form an initial stem point set;
converting the three-dimensional point cloud of the target plant from a global coordinate system to a local coordinate system;
extracting the highest point of each non-stem organ in the target plant according to the Z value of each non-stem organ point cloud in the local coordinate system, and respectively adding the highest point of each non-stem organ into the corresponding initial non-stem organ point set;
sequentially calculating the distance between each point to be segmented and each initial non-stem organ point set and the distance between each point to be segmented and the initial stem organ point set according to the Z value from large to small, selecting a target set corresponding to each point to be segmented from the initial stem point set and each initial non-stem organ point set, and adding each point to be segmented into the corresponding target set, wherein the points to be segmented are other points except the highest point of all stem point clouds and all non-stem organs;
And acquiring an optimal stem point set and each optimal non-stem organ point set, and acquiring stems and each non-stem organ in the target plant according to the optimal stem point set and each optimal non-stem point set.
Preferably, the step of dividing the stem point cloud in the three-dimensional point cloud of the target plant specifically includes:
acquiring top and bottom stems of the target three-dimensional plant, and adding the top and bottom stems into the initial stem set;
regarding a current seed point, taking all three-dimensional points in a first preset radius of the current seed point as current stem points, and adding all the current stem points into the initial stem point set;
calculating the growth direction of the current seed point according to a median normalized vector formed from all current stems to the current seed point and a normalized vector formed from the bottom stem to the top stem;
calculating the coordinates of the next seed point according to the coordinates of the current seed point, the first preset radius and the growth direction;
and calculating a projection point from the next seed point to a reference line, if the projection point does not exceed the top stem point, taking the next seed point as the current seed point again, repeating the above process until the projection point exceeds the top stem point, obtaining the initial stem point set, and enabling the reference line to pass through the top stem point and the bottom stem point, wherein the direction of the reference line is the same as the growth direction.
Preferably, the growing direction of the current seed point is calculated according to a median normalized vector formed from all current stems to the current seed point and a normalized vector formed from the bottom stem to the top stem, and a specific calculation formula is as follows:
wherein,indicating the growth direction of said current seed point, < >>Representing the median normalized vector formed from all current stalk points to said current seed point, +.>Representing a normalized vector formed between the bottom stem point to the top stem point 2 Represents L 2 The normal distance, mean, represents the median operation, a represents the set of all current shoot compositions, α represents a first preset weight, β represents a second preset weight, p A Represents any current shoot, s k Representing the current seed point, s n Representing the apical shoot, s 0 Representing the bottom end culm.
Preferably, the calculating the coordinates of the next seed point according to the coordinates of the current seed point, the first preset radius and the growth direction has the following specific calculation formula:
wherein s is k+1 Representing the coordinates of the next seed point, s k Representing the coordinates of the current seed point, R 1 Representing the first preset radius, v k Indicating the growth direction.
Preferably, the converting the three-dimensional point cloud of the target plant from a global coordinate system to a local coordinate system specifically includes:
taking the midpoints of all stem point clouds as the origin of the local coordinate system;
projecting all three-dimensional points in the global coordinate system onto a plane taking a Z axis of the local coordinate system as a normal vector, and acquiring a first principal component vector and a second principal component vector of projection of all three-dimensional point clouds through a principal component analysis method;
taking the first principal component vector as an X axis of the local coordinate system;
and taking the second principal component vector as a Y axis of the local coordinate system.
Preferably, the three-dimensional point cloud of the target plant is converted from a global coordinate system to a local coordinate system, and the specific calculation formula is as follows:
wherein (x ', y ', z ') represents three-dimensional point coordinates in the local coordinate system and (x, y, z) represents three-dimensional point coordinates in the global coordinate system, [ l ] 1 (x),l 1 (y),l 1 (z)]A unit direction vector, [ l ] representing the X-axis of the plant coordinate system in the global coordinate system 2 (x),l 2 (y),l 2 (z)]A unit direction vector, [ l ] representing the Y-axis of the plant coordinate system in the global coordinate system 3 (x),l 3 (y),l 3 (z)]A unit direction vector, [ O (x), O (y), O (Z) representing the Z-axis of the plant coordinate system in the global coordinate system ]An X coordinate value representing the local coordinate system origin under the global coordinate system.
Preferably, the calculating the distance between each point to be segmented and each initial non-stem organ point set and the distance between each point to be segmented and the initial stem organ point set sequentially according to the order of the Z values from large to small, and selecting the target set corresponding to each point to be segmented from the initial stem point set and each initial non-stem organ point set specifically includes:
for a current point to be segmented, calculating an average Euclidean distance between the current point to be segmented and each organ point set, wherein the organ point set comprises the initial stem point set and each initial non-stem organ point set;
selecting two organ point sets with the smallest average Euclidean distance as alternative point sets, adding the current point to be segmented into the initial stem point set if any alternative point set is the initial stem point set, otherwise, calculating the comprehensive distance between the current point to be segmented and each alternative point set, wherein the comprehensive distance consists of the average Euclidean distance and a local plane distance, the local plane distance represents the distance from the current point to be segmented to a local plane, and the local plane is generated by fitting a neighborhood point of the current point to be segmented within a second preset radius distance range;
And taking the candidate point set with smaller comprehensive distance as a target set corresponding to the current point to be segmented.
Preferably, the calculating the average euclidean distance from the current point to be segmented to each organ point set includes the following specific calculation formula:
wherein,representing the average Euclidean distance from the current point to be segmented to the mth organ point set, and p represents the current point to be segmented 2 Represents L 2 A paradigm distance, wherein A represents a K neighbor set of the current point to be segmented;
if the number of the m-th organ point concentration points is smaller than a preset value, taking the number of the m-th organ point concentration points as the value of K, otherwise, K is the preset value.
Preferably, the calculating the comprehensive distance between the current point to be segmented and each candidate point set has the following specific calculation formula:
f(x,y,z)＝n x x+n y y+n z z+d，
wherein Cm represents the comprehensive distance from the current point to be segmented to the mth organ point set,representing the average Euclidean distance from the current point to be segmented to the mth organ point set,/L>Representing the local plane distance from the current point to be segmented to the mth organ point set, f (x, y, z) representing the local plane, (n) x ,n y ,n z ) A normal vector representing the local plane, d representing the intercept of the local plane, (p) x ,p y ,p z ) Representing the coordinates of the current point to be segmented.
In a second aspect, an embodiment of the present invention provides a plant three-dimensional point cloud segmentation system, including:
the stem point cloud module is used for dividing stem point clouds in three-dimensional point clouds of target plants, wherein the three-dimensional point clouds of the target plants comprise stem point clouds and non-stem organ point clouds, and the divided stem point clouds form an initial stem point set;
the coordinate conversion module is used for converting the three-dimensional point cloud of the target plant from a global coordinate system to a local coordinate system;
the non-stem organ pre-segmentation module is used for extracting the highest point of each non-stem organ in the target plant according to the Z value of each non-stem organ point cloud in the local coordinate system, and respectively adding the highest point of each non-stem organ into the corresponding initial non-stem organ point set;
the organ segmentation module is used for sequentially calculating the distance between each point to be segmented and each initial non-stem organ point set and the distance between each point to be segmented and the initial stem organ point set according to the Z value from large to small, selecting a target set corresponding to each point to be segmented from the initial stem point set and each initial non-stem organ point set, and adding each point to be segmented into the corresponding target set, wherein the points to be segmented are other points except the highest point of all stem point clouds and all non-stem organs;
The optimal segmentation module is used for acquiring an optimal stem point set and each optimal non-stem organ point set, and acquiring stems and each non-stem organ in the target plant according to the optimal stem point set and each optimal non-stem point set.
In a third aspect, an embodiment of the present invention provides an electronic device, including a memory, a processor, and a computer program stored in the memory and capable of running on the processor, where the processor executes the program to implement the steps of the method for three-dimensional point cloud segmentation of a plant provided in the first aspect of the present invention.
In a fourth aspect, embodiments of the present invention provide a non-transitory computer readable storage medium having stored thereon a computer program which, when executed by a processor, implements the steps of the plant three-dimensional point cloud segmentation method provided in the first aspect of the present invention.
The three-dimensional point cloud segmentation method and system for the plants can realize three-dimensional point cloud segmentation of stalks, leaves, tassel and female spike organs in the plants; and can handle organ-dense scenes, i.e. can perform high-precision segmentation even when the organs are very close together or even close together. The embodiment of the invention greatly improves the segmentation precision of the corn point cloud organ and provides high-quality organ point cloud data for further corn phenotype detection and organ three-dimensional grid generation. In addition, the method can obtain high-quality corn point cloud labeling data and provide data support for three-dimensional deep learning of corn plants.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions of the prior art, the following description will briefly explain the drawings used in the embodiments or the description of the prior art, and it is obvious that the drawings in the following description are some embodiments of the present invention, and other drawings can be obtained according to these drawings without inventive effort for a person skilled in the art.
FIG. 1 is a flow chart of a method for dividing a three-dimensional point cloud of a plant according to an embodiment of the present invention;
FIG. 2 is a schematic representation of a three-dimensional point cloud, top and bottom stems of a maize plant provided in an embodiment of the present invention;
FIG. 3 is a schematic top-most view of a non-stem organ of a maize plant provided in an embodiment of the present invention;
fig. 4 is a schematic structural diagram of a plant three-dimensional point cloud segmentation system according to an embodiment of the present invention;
fig. 5 is a schematic entity structure diagram of an electronic device according to an embodiment of the present invention.
Detailed Description
For the purpose of making the objects, technical solutions and advantages of the embodiments of the present invention more apparent, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention, and it is apparent that the described embodiments are some embodiments of the present invention, but not all embodiments of the present invention. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
Fig. 1 is a flowchart of a method for dividing a three-dimensional point cloud of a plant according to an embodiment of the present invention, as shown in fig. 1, the method includes:
s1, segmenting a stem point cloud in a three-dimensional point cloud of a target plant, wherein the three-dimensional point cloud of the target plant comprises a stem point cloud and a non-stem organ point cloud, and the segmented stem point clouds form an initial stem point set;
for convenience of description, the embodiment of the invention uses corn as a target plant. Firstly, three-dimensional point cloud data of a corn plant are acquired, in general, the corn plant can be regarded as being composed of organs such as stems, leaves, tassels, female ears and the like, the leaves, the tassels, the female ears and the like are collectively called as non-stem organs, point clouds corresponding to the stems in the corn plant are called as stem point clouds, and point clouds corresponding to other non-stem organs are called as non-stem organ point clouds. The three-dimensional point cloud data of the corn plants are subjected to preliminary segmentation, point clouds corresponding to stem organs are extracted and are called stem clouds, and at the moment, a set formed by the stem clouds is called an initial stem set.
S2, converting the three-dimensional point cloud of the target plant from a global coordinate system to a local coordinate system;
in order to more conveniently determine the position of the point cloud in the plant through the coordinate values, the embodiment of the invention converts the three-dimensional point cloud coordinate of the plant from an original global coordinate system to a local coordinate system.
After the original three-dimensional point cloud coordinates are converted into the local coordinate system, the height of the point cloud on the plant can be judged through the coordinates of the Z value, and the larger the Z value is, the higher the height is. All subsequent operations are performed under a local coordinate system.
S3, extracting the highest point of each non-stem organ in the target plant according to the Z value of each non-stem organ point cloud in the local coordinate system, and respectively adding the highest point of each non-stem organ into the corresponding initial non-stem organ point set;
the highest point of each non-stalk organ is then extracted from all non-stalk organ point clouds as a seed point for that non-stalk organ point cloud for subsequent classification. Although the embodiment of the invention does not divide other non-stalk organs at this time, the highest point of each non-stalk organ can be estimated according to morphological characteristics of the corn organ.
The non-stem organ point clouds are other point clouds after the stem point clouds are removed by the three-dimensional point clouds of the corn plant, and the heights of all the point clouds can be represented by the Z value of the non-stem organ point clouds in the local coordinate system, so that in the embodiment of the invention, the highest point of each non-stem organ in the corn plant is extracted according to the Z value of each non-stem organ point cloud. And establishing an initial non-stalk organ point set for the highest point of each non-stalk organ, wherein the initial non-stalk organ point set is empty just at the beginning, and adding the highest point of each non-stalk organ into the corresponding initial non-stalk organ point set after obtaining the highest point of each non-stalk organ.
Specifically, in the embodiment of the invention, the extraction of the highest point of the non-stem organ is realized by the following method:
it was found by observation that there was only one peak for each non-stalk organ of maize, so the peak for the organ could be determined by finding the local maximum Z point in the point cloud. For any point p, searching for a neighborhood point within a sphere range with a certain radius, and if the Z value of the p point is larger than that of all the neighborhood points, the point is regarded as the highest point of a certain non-stem organ.
S4, sequentially calculating the distance between each point to be segmented and each initial non-stem organ point set and the distance between each point to be segmented and the initial stem organ point set according to the Z value from large to small, selecting a target set corresponding to each point to be segmented from the initial stem point set and each initial non-stem organ point set, and adding each point to be segmented into the corresponding target set, wherein the points to be segmented are other points except the highest point of all stem point clouds and all non-stem organs;
and after the highest points of the stem point clouds and the non-stem organs are segmented, taking the rest point clouds in the corn plants as points to be segmented. The Z values of the points to be segmented are sorted in the order from big to small, and calculated in the order from top to bottom.
Taking a certain point to be segmented as an example for illustration, calculating the distance between the point to be segmented and each initial non-stem organ point set and the distance between the point to be segmented and the initial stem point set, selecting a target set corresponding to the point to be segmented from the distance between the point to be segmented and each point cloud set, and adding the point to be segmented into the target set.
And carrying out corresponding operation on each point to be segmented until each point to be segmented is distributed.
S5, acquiring an optimal stem point set and each optimal non-stem organ point set, and acquiring stems and each non-stem organ in the target plant according to the optimal stem point set and each optimal non-stem point set.
After the steps, taking the initial stalk point set as an optimal stalk point set, taking the initial non-stalk organ point set as an optimal non-stalk organ point set, and obtaining the stalk and each non-stalk organ in the corn plant according to the optimal stalk point set and each optimal non-stalk organ point set.
The three-dimensional point cloud segmentation method for the plants can realize three-dimensional point cloud segmentation of stalks, leaves, tassel and female ear organs in the plants; and can process scenes with dense organs, i.e. organs very close together or even close together, and can also perform high-precision segmentation. The embodiment of the invention greatly improves the segmentation precision of the corn point cloud organ and provides high-quality organ point cloud data for further corn phenotype detection and organ three-dimensional grid generation. In addition, the method can obtain high-quality corn point cloud labeling data and provide data support for three-dimensional deep learning of corn plants.
On the basis of the foregoing embodiment, preferably, the segmenting the stem point cloud in the three-dimensional point cloud of the target plant specifically includes:
acquiring top and bottom stems of the target three-dimensional plant, and adding the top and bottom stems into the initial stem set;
FIG. 2 is a schematic view of three-dimensional point clouds, top and bottom stems of a maize plant provided in an embodiment of the present invention, as shown in FIG. 2, S in the drawing 0 Represents the bottom end stem point, S n Representing the apical shoot.
Firstly, the top stem point and the bottom stem point of the corn plant are extracted, the extraction method can be used for extracting through a preset algorithm, and the corn plant can be marked through a manual means, and the embodiment of the invention is not particularly limited.
In addition, these two points are added to the initial set of stems, with the original initial set of stems being empty.
After the bottom stem point and the top stem point are selected, the bottom stem point is used as the current seed point for carrying out the stem point cloud segmentation in the initial stage.
Regarding a current seed point, taking all three-dimensional points in a first preset radius of the current seed point as current stem points, and adding all the current stem points into the initial stem point set;
calculating the growth direction of the current seed point according to a median normalized vector formed from all current stems to the current seed point and a normalized vector formed from the bottom stem to the top stem;
Calculating the coordinates of the next seed point according to the coordinates of the current seed point, the first preset radius and the growth direction;
and calculating a projection point from the next seed point to a reference line, if the projection point does not exceed the top stem point, taking the next seed point as the current seed point again, repeating the above process until the projection point exceeds the top stem point, obtaining the initial stem point set, and enabling the reference line to pass through the top stem point and the bottom stem point, wherein the direction of the reference line is the same as the growth direction.
The segmentation algorithm is an iterative process, and the current seed point of the current iteration is s under the assumption that the current algorithm is in the kth iteration k The stem segmentation process is as follows:
in s k The set of points a within a first preset radius (R1) is divided into current stems, with the R1 parameter value set by the user. Then calculate each point in the set of points A and s k Unit direction vectors among the unit vectors are calculated, and then the median vector of the unit vectors is obtained and normalized; and then calculating a normalized vector formed between the bottom stem point and the top stem point, and obtaining the growth direction of the current seed point according to the calculated two normalized vectors. The normalization vector between the bottom stem point and the top stem point has a correcting function, so that the growth direction of the whole region is ensured not to deviate from the growth direction of the stem seriously, and the stem can be accurately segmented.
And according to the coordinates of the current seed point, the first preset radius and the growth direction, obtaining the coordinates of the next seed point for the next iteration, calculating the projection point from the next seed point to the datum line, if the projection point from the next seed point to the datum line does not exceed the top stem point, indicating that the next seed point is still on the stem, taking the next seed point as the current seed point again, repeating the above process until the projection point from the next seed point to the datum line exceeds the top stem point, indicating that the next seed point is no longer on the stem, and ending the iteration process.
In the embodiment of the invention, the datum line is a straight line, and the direction of the straight line is the same as the growth direction of the current seed point and passes through the top stem point and the bottom stem point.
On the basis of the above embodiment, preferably, the growth direction of the current seed point is calculated according to a median normalized vector formed from all current stems to the current seed point and a normalized vector formed from the bottom stem to the top stem, and a specific calculation formula is as follows:
wherein,indicating the growth direction of said current seed point, < >>Representing the median normalized vector formed from all current stalk points to said current seed point, +. >Representing a normalized vector formed between the bottom stem point to the top stem point 2 Represents L 2 The normal distance, mean, represents the median operation, a represents the set of all current shoot compositions, α represents a first preset weight, β represents a second preset weight, p A Represents any current shoot, s k Representing the current seed point, s n Representing the apical shoot, s 0 Representing the bottom end culm.
In the current implementation, α=0.2 and β=0.8 are taken in the embodiment of the present invention, and under this parameter, it can be ensured that the whole segmentation process can be correctly segmented to the top of the stem along the bottom of the stem at different first preset radius values.
On the basis of the foregoing embodiment, preferably, the calculating the coordinates of the next seed point according to the coordinates of the current seed point, the first preset radius and the growth direction has the following specific calculation formula:
wherein s is k+1 Representing the coordinates of the next seed point, s k Representing the coordinates of the current seed point, R 1 Representing the first preset radius, v k Indicating the growth direction.
On the basis of the foregoing embodiment, preferably, the converting the three-dimensional point cloud in the target plant from the global coordinate system to the local coordinate system specifically includes:
Taking the midpoints of all stem point clouds as the origin of the local coordinate system;
projecting all three-dimensional point clouds in the global coordinate system onto a plane taking a Z axis of the local coordinate system as a normal vector, and acquiring a first principal component vector and a second principal component vector of projection of all the three-dimensional point clouds by a principal component analysis method;
taking the first principal component vector as an X axis of the local coordinate system;
and taking the second principal component vector as a Y axis of the local coordinate system.
In the embodiment of the invention, the local coordinate system takes the midpoint O of all the stem point clouds in the initial stem point set obtained in the step S1 as an origin, and is composed of three mutually perpendicular unit vectors, wherein the Z axis is the central axis l of all the stem point clouds 3 In the embodiment of the invention, the method is obtained by fitting by a least square method.
Then all three-dimensional point clouds are projected onto a plane taking a Z axis of a local coordinate system as a normal vector, and a principal component analysis method is adopted to calculate a first principal component vector l of the projected points 1 And a second principal component vector l 2 ，l 1 Is the X axis of a local coordinate system, l 2 Is the Y-axis of the local coordinate system, then any point (x, Y, z) in the initial global coordinate system is transferred to the coordinates (x ', Y ', z ') in the local coordinate system by the following equation.
On the basis of the foregoing embodiment, preferably, the calculating, in order from large to small, the distance between each point to be segmented and each initial non-stem organ point set and the distance between each point to be segmented and the initial stem organ point set sequentially, and selecting, from the initial stem point set and each initial non-stem organ point set, a target set corresponding to each point to be segmented specifically includes:
for a current point to be segmented, calculating an average Euclidean distance between the current point to be segmented and each organ point set, wherein the organ point set comprises the initial stem point set and each initial non-stem organ point set;
specifically, a certain point to be segmented is taken as a current point to be segmented, and an average Euclidean distance between the current point to be segmented and an initial stem point set and an average Euclidean distance between the current point to be segmented and each initial non-stem organ point set are calculated.
And selecting two organ point sets with the smallest average Euclidean distance as alternative point sets, adding the current point to be segmented into the initial stem point set if any alternative point set is the initial stem point set, otherwise, calculating the comprehensive distance between the current point to be segmented and each alternative point set, wherein the comprehensive distance consists of the average Euclidean distance and a local plane distance, the local plane distance represents the distance from the current point to be segmented to a local plane, and the local plane is generated by fitting a neighborhood point of the current point to be segmented within a second preset radius distance range.
And taking the two organ point sets with the smallest average Euclidean distance as alternative point sets, if one of the two alternative point sets is an initial stem point set, taking the target set of the current point to be segmented as the initial stem point set, and adding the current point to be segmented into the initial stem point set.
If the two alternative point sets are non-stem organ point sets, calculating the comprehensive distance between the current point to be segmented and the two alternative point sets, and taking the alternative point set with the small comprehensive distance as a target set corresponding to the current point to be segmented.
In the embodiment of the invention, the comprehensive distance consists of two parts, one part is an average Euclidean distance, the other part is a local plane distance, the local plane distance represents the distance from the current point to be segmented to the local plane, and the local plane is formed by fitting the field points of the current point to be segmented within a second preset radius distance range.
And taking the candidate point set with smaller comprehensive distance as a target set corresponding to the current point to be segmented.
And then taking the alternative point set with smaller comprehensive distance as a target set corresponding to the current point to be segmented.
On the basis of the foregoing embodiment, preferably, the calculating the average euclidean distance between the current point to be segmented and each organ point set includes the following specific calculation formula:
Wherein,representing the average Euclidean distance from the current point to be segmented to the mth organ point set, and p represents the current point to be segmented 2 Represents L 2 A paradigm distance, wherein A represents a K neighbor set of the current point to be segmented;
if the number of the point clouds in the mth organ point set is smaller than a preset value, taking the number of the point clouds in the mth organ point set as the value of K, otherwise, K is the preset value.
On the basis of the foregoing embodiment, preferably, the calculating the comprehensive distance between the current point to be segmented and each candidate point set includes the following specific calculation formula:
f(x,y,z)＝n x x+n y y+n z z+d，
wherein Cm represents the comprehensive distance from the current point to be segmented to the mth organ point set,representing the average Euclidean distance from the current point to be segmented to the mth organ point set,/L>Representing the local plane distance from the current point to be segmented to the mth organ point set, f (x, y, z) representing the local plane, (n) x ,n y ,n z ) A normal vector representing the local plane, d representing the intercept of the local plane, (p) x ,p y ,p z ) Representing the coordinates of the current point to be segmented.
The organ segmentation algorithm of the embodiment of the invention can be specifically expressed as follows: FIG. 3 is a schematic diagram showing the highest points of non-stem organs of a maize plant according to the embodiment of the present invention, as shown in FIG. 3, assuming that the highest points of n non-stem organs are obtained in total, wherein the jth highest point uses p j (j=1, 2, …, n), then the n highest points are separated from φ u And respectively placing the initial non-stem organ points in n initial non-stem organ point sets, wherein the placing method comprises the following steps of: will p j Point placement onto organ collectionsIn (I)>Representing the initial set of stems.
The points to be segmented are classified sequentially from top to bottom, namely, the points to be segmented with larger Z coordinate values are classified first. In specific implementation, the classification process of the embodiment of the invention is as follows:
1) Will phi u The points in (2) are ordered in order from large to small according to Z values, and then are sequentially ordered according to the orderAnd taking out the point, and executing the following steps.
2) From phi u Taking out a point p, judging which organ the point belongs to, and setting the point p to the m-th organ point setAverage Euclidean distance>Calculating the average Euclidean distance of point p to all organ point sets +.>From which the two organs with the smallest values are extractedAnd->As an alternative set of points. If one of the two alternative points represents a stalk organ, then point p is segmented directly into the stalk organ, step 3) is performed.
If both candidate point sets are non-stem organs, inBased on the distance, a local plane distance +.>Obtaining the comprehensive distance C m Calculating points p to ∈>And->The integrated distance between them, the set with the smallest integrated distance as the target set of points p, by +. >Representing the target set of points p.
3) The point p is from phi u Move toUpdated->To be used in the sorting operation of the next point if phi u If the operation is empty, the whole classifying operation is finished, otherwise, the step 2) is continuously executed.
The embodiment of the invention can realize three-dimensional point cloud segmentation of stalks, leaves, tassel and female spike organs in plant plants; and can process scenes with dense organs, i.e. organs very close together or even close together, and can also perform high-precision segmentation.
Fig. 4 is a schematic structural diagram of a plant three-dimensional point cloud segmentation system according to an embodiment of the present invention, where the system includes: a stalk point cloud module 401, a coordinate conversion module 402, a non-stalk organ pre-segmentation module 403, an organ segmentation module 404, and an optimal segmentation module 405, wherein:
the stem point cloud module 401 is configured to segment stem point clouds in a three-dimensional point cloud of a target plant, where the three-dimensional point cloud of the target plant includes a stem point cloud and a non-stem organ point cloud, and the segmented stem point clouds form an initial stem point set;
the coordinate conversion module 402 is used for converting the three-dimensional point cloud of the target plant from a global coordinate system to a local coordinate system;
the non-stalk organ pre-segmentation module 403 is configured to extract a highest point of each non-stalk organ in the target plant according to a Z value of each non-stalk organ point cloud in the local coordinate system, and add the highest point of each non-stalk organ to a corresponding initial non-stalk organ point set respectively;
The organ segmentation module 404 is configured to sequentially calculate, according to the order of the Z value from large to small, a distance between each point to be segmented and each initial non-stalk organ point set and a distance between each point to be segmented and the initial stalk organ point set, select a target set corresponding to each point to be segmented from the initial stalk point set and each initial non-stalk organ point set, and add each point to be segmented into the corresponding target set, where the point to be segmented is other points except the highest point of all stalk point clouds and all non-stalk organs;
the optimal segmentation module 405 is configured to obtain an optimal shoot set and each optimal non-stem organ set, and obtain a stem and each non-stem organ in the target plant according to the optimal shoot set and each optimal non-stem set.
The embodiment of the device provided by the embodiment of the present invention is for implementing the above embodiments of the method, and specific flow and details refer to the above embodiments of the method, which are not repeated herein.
Fig. 5 is a schematic physical structure diagram of an electronic device according to an embodiment of the present invention, where, as shown in fig. 5, the electronic device may include: a processor (processor) 501, a communication interface (Communications Interface) 502, a memory (memory) 503 and a bus 504, wherein the processor 501, the communication interface 502, and the memory 503 communicate with each other via the bus 504. The communication interface 502 may be used for information transfer of an electronic device. The processor 501 may invoke logic instructions in the memory 503 to perform a method comprising:
Dividing a stem point cloud in a three-dimensional point cloud of a target plant, wherein the three-dimensional point cloud of the target plant comprises a stem point cloud and a non-stem organ point cloud, and the divided stem point clouds form an initial stem point set;
converting the three-dimensional point cloud of the target plant from a global coordinate system to a local coordinate system;
extracting the highest point of each non-stem organ in the target plant according to the Z value of each non-stem organ point cloud in the local coordinate system, and respectively adding the highest point of each non-stem organ into the corresponding initial non-stem organ point set;
sequentially calculating the distance between each point to be segmented and each initial non-stem organ point set and the distance between each point to be segmented and the initial stem organ point set according to the Z value from large to small, selecting a target set corresponding to each point to be segmented from the initial stem point set and each initial non-stem organ point set, and adding each point to be segmented into the corresponding target set, wherein the points to be segmented are other points except the highest point of all stem point clouds and all non-stem organs;
and acquiring an optimal stem point set and each optimal non-stem organ point set, and acquiring stems and each non-stem organ in the target plant according to the optimal stem point set and each optimal non-stem point set.
Further, the logic instructions in the memory 503 described above may be implemented in the form of software functional units and may be stored in a computer readable storage medium when sold or used as a stand alone product. Based on this understanding, the technical solution of the present invention may be embodied essentially or in a part contributing to the prior art or in the form of a software product stored in a storage medium, comprising several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the above-described method embodiments of the present invention. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a random access Memory (RAM, random Access Memory), a magnetic disk, or an optical disk, or other various media capable of storing program codes.
In another aspect, embodiments of the present invention further provide a non-transitory computer readable storage medium having stored thereon a computer program, which when executed by a processor is implemented to perform the transmission method provided in the above embodiments, for example, including:
Dividing a stem point cloud in a three-dimensional point cloud of a target plant, wherein the three-dimensional point cloud of the target plant comprises a stem point cloud and a non-stem organ point cloud, and the divided stem point clouds form an initial stem point set;
converting the three-dimensional point cloud of the target plant from a global coordinate system to a local coordinate system;
extracting the highest point of each non-stem organ in the target plant according to the Z value of each non-stem organ point cloud in the local coordinate system, and respectively adding the highest point of each non-stem organ into the corresponding initial non-stem organ point set;
sequentially calculating the distance between each point to be segmented and each initial non-stem organ point set and the distance between each point to be segmented and the initial stem organ point set according to the Z value from large to small, selecting a target set corresponding to each point to be segmented from the initial stem point set and each initial non-stem organ point set, and adding each point to be segmented into the corresponding target set, wherein the points to be segmented are other points except the highest point of all stem point clouds and all non-stem organs;
and acquiring an optimal stem point set and each optimal non-stem organ point set, and acquiring stems and each non-stem organ in the target plant according to the optimal stem point set and each optimal non-stem point set.
The apparatus embodiments described above are merely illustrative, wherein the elements illustrated as separate elements may or may not be physically separate, and the elements shown as elements may or may not be physical elements, may be located in one place, or may be distributed over a plurality of network elements. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment. Those of ordinary skill in the art will understand and implement the present invention without undue burden.
From the above description of the embodiments, it will be apparent to those skilled in the art that the embodiments may be implemented by means of software plus necessary general hardware platforms, or of course may be implemented by means of hardware. Based on this understanding, the foregoing technical solution may be embodied essentially or in a part contributing to the prior art in the form of a software product, which may be stored in a computer readable storage medium, such as ROM/RAM, a magnetic disk, an optical disk, etc., including several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute the method described in the respective embodiments or some parts of the embodiments.
Finally, it should be noted that: the above embodiments are only for illustrating the technical solution of the present invention, and are not limiting; although the invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical scheme described in the foregoing embodiments can be modified or some technical features thereof can be replaced by equivalents; such modifications and substitutions do not depart from the spirit and scope of the technical solutions of the embodiments of the present invention.
Claims (7)
1. The plant three-dimensional point cloud segmentation method is characterized by comprising the following steps of:
dividing a stem point cloud in a three-dimensional point cloud of a target plant, wherein the three-dimensional point cloud of the target plant comprises a stem point cloud and a non-stem organ point cloud, and the divided stem point clouds form an initial stem point set;
converting the three-dimensional point cloud of the target plant from a global coordinate system to a local coordinate system;
extracting the highest point of each non-stem organ in the target plant according to the Z value of each non-stem organ point cloud in the local coordinate system, and respectively adding the highest point of each non-stem organ into the corresponding initial non-stem organ point set;
sequentially calculating the distance between each point to be segmented and each initial non-stem organ point set and the distance between each point to be segmented and the initial stem organ point set according to the Z value from large to small, selecting a target set corresponding to each point to be segmented from the initial stem point set and each initial non-stem organ point set, and adding each point to be segmented into the corresponding target set, wherein the points to be segmented are other points except the highest point of all stem point clouds and all non-stem organs;
Acquiring an optimal stem point set and each optimal non-stem organ point set, and acquiring stems and each non-stem organ in the target plant according to the optimal stem point set and each optimal non-stem organ point set;
the converting the three-dimensional point cloud of the target plant from a global coordinate system to a local coordinate system specifically comprises the following steps:
taking the midpoints of all stem point clouds as the origin of the local coordinate system;
projecting all three-dimensional points in the global coordinate system onto a plane taking a Z axis of the local coordinate system as a normal vector, and acquiring a first principal component vector and a second principal component vector of projection of all three-dimensional point clouds through a principal component analysis method;
taking the first principal component vector as an X axis of the local coordinate system;
taking the second principal component vector as a Y axis of the local coordinate system;
extracting the highest point of each non-stem organ in the target plant according to the Z value of the point cloud of each non-stem organ in the local coordinate system, and respectively adding the highest point of each non-stem organ into the corresponding initial non-stem organ point set, wherein the method specifically comprises the following steps of:
searching for non-stalk organ point clouds in all adjacent domains of a certain radius sphere range of the non-stalk organ point clouds, when the Z value of the non-stalk organ point clouds is larger than the Z value of the non-stalk organ point clouds in all the adjacent domains, determining the non-stalk organ point clouds as the highest point of each non-stalk organ in the target plant, and respectively adding the highest point of each non-stalk organ into the corresponding initial non-stalk organ point set;
The method comprises the steps of sequentially calculating the distance between each point to be segmented and each initial non-stem organ point set and the distance between each point to be segmented and the initial stem point set according to the Z value from large to small, and selecting a target set corresponding to each point to be segmented from the initial stem point set and each initial non-stem organ point set, wherein the method specifically comprises the following steps:
for a current point to be segmented, calculating an average Euclidean distance between the current point to be segmented and each organ point set, wherein the organ point set comprises the initial stem point set and each initial non-stem organ point set;
selecting two organ point sets with the smallest average Euclidean distance as alternative point sets, adding the current point to be segmented into the initial stem point set if any alternative point set is the initial stem point set, otherwise, calculating the comprehensive distance between the current point to be segmented and each alternative point set, wherein the comprehensive distance consists of the average Euclidean distance and a local plane distance, the local plane distance represents the distance from the current point to be segmented to a local plane, and the local plane is generated by fitting a neighborhood point of the current point to be segmented within a second preset radius distance range;
Taking the alternative point set with smaller comprehensive distance as a target set corresponding to the current point to be segmented;
the comprehensive distance between the current point to be segmented and each candidate point set is calculated, and a specific calculation formula is as follows:
f(x,y,z)＝n x x+n y y+n z z+d，
wherein C is m Representing the comprehensive distance from the current point to be segmented to the mth organ point set,representing the average Euclidean distance from the current point to be segmented to the mth organ point set,/L>Representing the local plane distance from the current point to be segmented to the mth organ point set, f (x, y, z) representing the local plane, (n) x ,n y ,n z ) A normal vector representing the local plane, d representing the intercept of the local plane, (p) x ,p y ,p z ) Representing the coordinates of the current point to be segmented;
the obtaining an optimal stem point set and each optimal non-stem organ point set, according to the optimal stem point set and each optimal non-stem organ point set, obtaining stems and each non-stem organ in the target plant specifically includes:
adding each point to be segmented into the corresponding initial stalk point set to obtain a stalk point set serving as an optimal stalk point set, adding the highest point of each non-stalk organ into the corresponding initial non-stalk organ point set, adding each point to be segmented into the corresponding initial non-stalk organ point set, obtaining the initial non-stalk organ point set serving as an optimal non-stalk organ point set, and obtaining the stalk and each non-stalk organ in a plant according to the optimal stalk point set and each optimal non-stalk organ point set.
2. The method for dividing the three-dimensional point cloud of the plant according to claim 1, wherein the step of dividing the stem point cloud of the three-dimensional point cloud of the target plant specifically comprises the steps of:
acquiring top and bottom stems of the three-dimensional point cloud of the target plant, and adding the top and bottom stems into the initial stem set;
regarding a current seed point, taking all three-dimensional points in a first preset radius of the current seed point as current stem points, and adding all the current stem points into the initial stem point set;
calculating the growth direction of the current seed point according to a median normalized vector formed from all current stems to the current seed point and a normalized vector formed from the bottom stem to the top stem;
calculating the coordinates of the next seed point according to the coordinates of the current seed point, the first preset radius and the growth direction;
and calculating a projection point from the next seed point to a reference line, if the projection point does not exceed the top stem point, taking the next seed point as the current seed point again, repeating the above process until the projection point exceeds the top stem point, obtaining the initial stem point set, and enabling the reference line to pass through the top stem point and the bottom stem point, wherein the direction of the reference line is the same as the growth direction.
3. The plant three-dimensional point cloud segmentation method according to claim 2, wherein the growth direction of the current seed point is calculated according to a median normalized vector formed from all current stems to the current seed point and a normalized vector formed from the bottom stem point to the top stem point, and a specific calculation formula is as follows:
wherein,indicating the growth direction of said current seed point, < >>Representing the median normalized vector formed from all current stalk points to said current seed point, +.>Representing a normalized vector formed between the bottom stem point to the top stem point 2 Represents L 2 The normal distance, mean, represents the median operation, a represents the set of all current shoot compositions, α represents a first preset weight, β represents a second preset weight, p A Represents any current shoot, s k Representing the current seed point, s n Representing the apical shoot, s 0 Representing the bottom end culm.
4. The method for three-dimensional point cloud segmentation of plants according to claim 2, wherein the coordinates of the next seed point are calculated according to the coordinates of the current seed point, the first preset radius and the growth direction, and a specific calculation formula is as follows:
Wherein s is k+1 Representing the coordinates of the next seed point, s k Representing the coordinates of the current seed point, R 1 Representing the first preset radius, v k Indicating the growth direction.
5. The method for three-dimensional point cloud segmentation of plants according to claim 1, wherein the target plant three-dimensional point cloud is transformed from a global coordinate system to a local coordinate system according to the following specific calculation formula:
wherein (x ', y ', z ') represents three-dimensional point coordinates in the local coordinate system and (x, y, z) represents three-dimensional point coordinates in the global coordinate system, [ l ] 1 (x),l 1 (y),l 1 (z)]A unit direction vector, [ l ] representing the X-axis of the plant coordinate system in the global coordinate system 2 (x),l 2 (y),l 2 (z)]A unit direction vector, [ l ] representing the Y-axis of the plant coordinate system in the global coordinate system 3 (x),l 3 (y),l 3 (z)]A unit direction vector, [ O (x), O (y), O (Z) representing the Z-axis of the plant coordinate system in the global coordinate system]An X coordinate value representing the local coordinate system origin under the global coordinate system.
6. The method for three-dimensional point cloud segmentation of plants according to claim 1, wherein the average euclidean distance from the current point to be segmented to each organ point set is calculated according to the following specific calculation formula:
Wherein,representing the average Euclidean distance from the current point to be segmented to the mth organ point set, and p represents the current point to be segmented 2 Represents L 2 A paradigm distance, wherein A represents a K neighbor set of the current point to be segmented;
if the number of the m-th organ point concentration points is smaller than a preset value, taking the number of the m-th organ point concentration points as the value of K, otherwise, K is the preset value.
7. A plant three-dimensional point cloud segmentation system, comprising:
the stem point cloud module is used for dividing stem point clouds in three-dimensional point clouds of target plants, wherein the three-dimensional point clouds of the target plants comprise stem point clouds and non-stem organ point clouds, and the divided stem point clouds form an initial stem point set;
the coordinate conversion module is used for converting the three-dimensional point cloud of the target plant from a global coordinate system to a local coordinate system;
the non-stem organ pre-segmentation module is used for extracting the highest point of each non-stem organ in the target plant according to the Z value of each non-stem organ point cloud in the local coordinate system, and respectively adding the highest point of each non-stem organ into the corresponding initial non-stem organ point set;
the organ segmentation module is used for sequentially calculating the distance between each point to be segmented and each initial non-stem organ point set and the distance between each point to be segmented and the initial stem organ point set according to the Z value from large to small, selecting a target set corresponding to each point to be segmented from the initial stem point set and each initial non-stem organ point set, and adding each point to be segmented into the corresponding target set, wherein the points to be segmented are other points except the highest point of all stem point clouds and all non-stem organs;
The optimal segmentation module is used for acquiring an optimal stem point set and each optimal non-stem organ point set, and acquiring stems and each non-stem organ in the target plant according to the optimal stem point set and each optimal non-stem organ point set;
the coordinate conversion module is specifically used for:
taking the midpoints of all stem point clouds as the origin of the local coordinate system;
projecting all three-dimensional points in the global coordinate system onto a plane taking a Z axis of the local coordinate system as a normal vector, and acquiring a first principal component vector and a second principal component vector of projection of all three-dimensional point clouds through a principal component analysis method;
taking the first principal component vector as an X axis of the local coordinate system;
taking the second principal component vector as a Y axis of the local coordinate system;
the non-stem organ pre-segmentation module is specifically used for:
searching for non-stalk organ point clouds in all adjacent domains of a certain radius sphere range of the non-stalk organ point clouds, when the Z value of the non-stalk organ point clouds is larger than the Z value of the non-stalk organ point clouds in all the adjacent domains, determining the non-stalk organ point clouds as the highest point of each non-stalk organ in the target plant, and respectively adding the highest point of each non-stalk organ into the corresponding initial non-stalk organ point set;
The organ segmentation module is specifically configured to:
for a current point to be segmented, calculating an average Euclidean distance between the current point to be segmented and each organ point set, wherein the organ point set comprises the initial stem point set and each initial non-stem organ point set;
selecting two organ point sets with the smallest average Euclidean distance as alternative point sets, adding the current point to be segmented into the initial stem point set if any alternative point set is the initial stem point set, otherwise, calculating the comprehensive distance between the current point to be segmented and each alternative point set, wherein the comprehensive distance consists of the average Euclidean distance and a local plane distance, the local plane distance represents the distance from the current point to be segmented to a local plane, and the local plane is generated by fitting a neighborhood point of the current point to be segmented within a second preset radius distance range;
taking the alternative point set with smaller comprehensive distance as a target set corresponding to the current point to be segmented;
the comprehensive distance between the current point to be segmented and each candidate point set is calculated, and a specific calculation formula is as follows:
f(x,y,z)＝n x x+n y y+n z z+d，
wherein C is m Representing the comprehensive distance from the current point to be segmented to the mth organ point set,representing the average Euclidean distance from the current point to be segmented to the mth organ point set,/L >Representing the local plane distance from the current point to be segmented to the mth organ point set, f (x, y, z) representing the local plane, (n) x ,n y ,n z ) A normal vector representing the local plane, d representing the intercept of the local plane, (p) x ,p y ,p z ) Representing the coordinates of the current point to be segmented;
the optimal segmentation module is specifically configured to:
adding each point to be segmented into the corresponding initial stalk point set to obtain a stalk point set serving as an optimal stalk point set, adding the highest point of each non-stalk organ into the corresponding initial non-stalk organ point set, adding each point to be segmented into the corresponding initial non-stalk organ point set, obtaining the initial non-stalk organ point set serving as an optimal non-stalk organ point set, and obtaining the stalk and each non-stalk organ in a plant according to the optimal stalk point set and each optimal non-stalk organ point set.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010475872.9A CN111768413B (en) | 2020-05-29 | 2020-05-29 | Plant three-dimensional point cloud segmentation method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010475872.9A CN111768413B (en) | 2020-05-29 | 2020-05-29 | Plant three-dimensional point cloud segmentation method and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111768413A CN111768413A (en) | 2020-10-13 |
| CN111768413B true CN111768413B (en) | 2023-12-05 |
Family
ID=72720640
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010475872.9A Active CN111768413B (en) | 2020-05-29 | 2020-05-29 | Plant three-dimensional point cloud segmentation method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111768413B (en) |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103065352A (en) * | 2012-12-20 | 2013-04-24 | 北京农业信息技术研究中心 | Plant three-dimensional reconstruction method based on image and scanning data |
| CN103218817A (en) * | 2013-04-19 | 2013-07-24 | 深圳先进技术研究院 | Partition method and partition system of plant organ point clouds |
| CN104484873A (en) * | 2014-11-28 | 2015-04-01 | 浙江大学 | Method for separating leaves and stem in three-dimensional point cloud of plant |
| CN105354829A (en) * | 2015-10-08 | 2016-02-24 | 西北农林科技大学 | Self-adaptive point cloud data segmenting method |
| CN109035286A (en) * | 2018-06-29 | 2018-12-18 | 北京农业信息技术研究中心 | Plant three dimensional point cloud high throughput acquisition methods and device |
| CN109708578A (en) * | 2019-02-25 | 2019-05-03 | 中国农业科学院农业信息研究所 | A kind of plant phenotype parameter measuring apparatus, method and system |
| CN109887020A (en) * | 2019-02-25 | 2019-06-14 | 中国农业科学院农业信息研究所 | A kind of plant organ's separation method and system |
| CN110853044A (en) * | 2019-04-25 | 2020-02-28 | 华中农业大学 | Potted corn point cloud rapid segmentation method based on conditional Euclidean clustering |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US10520482B2 (en) * | 2012-06-01 | 2019-12-31 | Agerpoint, Inc. | Systems and methods for monitoring agricultural products |
| US11188752B2 (en) * | 2018-03-08 | 2021-11-30 | Regents Of The University Of Minnesota | Crop biometrics detection |
| US10891482B2 (en) * | 2018-07-10 | 2021-01-12 | Adroit Robotics | Systems, devices, and methods for in-field diagnosis of growth stage and crop yield estimation in a plant area |
-
2020
- 2020-05-29 CN CN202010475872.9A patent/CN111768413B/en active Active
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103065352A (en) * | 2012-12-20 | 2013-04-24 | 北京农业信息技术研究中心 | Plant three-dimensional reconstruction method based on image and scanning data |
| CN103218817A (en) * | 2013-04-19 | 2013-07-24 | 深圳先进技术研究院 | Partition method and partition system of plant organ point clouds |
| CN104484873A (en) * | 2014-11-28 | 2015-04-01 | 浙江大学 | Method for separating leaves and stem in three-dimensional point cloud of plant |
| CN105354829A (en) * | 2015-10-08 | 2016-02-24 | 西北农林科技大学 | Self-adaptive point cloud data segmenting method |
| CN109035286A (en) * | 2018-06-29 | 2018-12-18 | 北京农业信息技术研究中心 | Plant three dimensional point cloud high throughput acquisition methods and device |
| CN109708578A (en) * | 2019-02-25 | 2019-05-03 | 中国农业科学院农业信息研究所 | A kind of plant phenotype parameter measuring apparatus, method and system |
| CN109887020A (en) * | 2019-02-25 | 2019-06-14 | 中国农业科学院农业信息研究所 | A kind of plant organ's separation method and system |
| AU2019100886A4 (en) * | 2019-02-25 | 2019-09-12 | Agricultural Information Institute Of Caas | Plant organ image separation method and system |
| CN110853044A (en) * | 2019-04-25 | 2020-02-28 | 华中农业大学 | Potted corn point cloud rapid segmentation method based on conditional Euclidean clustering |
Non-Patent Citations (7)
| Title |
|---|
| Automated classification of stems and leaves of potted plants based on point cloud data.arXiv:2002.12536.2020,1-31. * |
| MVS-Pheno: A Portable and Low-Cost Phenotyping Platform for Maize Shoots Using Multiview Stereo 3D Reconstruction;Sheng Wu等;Plant Phenomics;1-8 * |
| Point cloud based iterative segmentation technique for 3D plant phenotyping;J. Liu 等;2018 IEEE International Conference on Information and Automation (ICIA);1072-1077 * |
| Stem–Leaf Segmentation and Phenotypic Trait Extraction of Individual Maize Using Terrestrial LiDAR Data;S. Jin 等;IEEE Transactions on Geoscience and Remote Sensing;第57卷(第3期);1336-1346 * |
| 基于三维数字化的玉米株型参数提取方法研究;温维亮 等;中国农业科学;第51卷(第6期);1034-1044 * |
| 水稻植株三维重建技术研究;吴丹;中国硕士学位论文全文数据库 (农业科技辑)(第（2020）02期);D047-16 * |
| 玉米株型冠层三维数字化与结构解析技术研究;温维亮;中国博士学位论文全文数据库 (农业科技辑)(第(2018)07期);D047-20 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111768413A (en) | 2020-10-13 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111724433B (en) | Crop phenotype parameter extraction method and system based on multi-view vision | |
| CN109146948B (en) | Crop growth phenotype parameter quantification and yield correlation analysis method based on vision | |
| CN108198244B (en) | Apple leaf point cloud simplification method and device | |
| CN108681692B (en) | Method for identifying newly added buildings in remote sensing image based on deep learning | |
| CN106340016B (en) | A kind of DNA quantitative analysis method based on microcytoscope image | |
| CN110232685A (en) | Space pelvis parameter auto-testing method based on deep learning | |
| CN109993750A (en) | A kind of segmentation recognition method and system, terminal and readable storage medium storing program for executing of hand jnjuries | |
| CN111429490A (en) | Agricultural and forestry crop three-dimensional point cloud registration method based on calibration ball | |
| CN110322497A (en) | A kind of interactive point cloud object extraction method based on three-dimensional visualization | |
| CN110246218A (en) | Method for reconstructing, the space pelvis measurement method of parameters of femoral head threedimensional model | |
| CN111291686B (en) | Extraction method and system for crop root-fruit phenotype parameters and root-fruit phenotype discrimination method and system | |
| CN116452648B (en) | Point cloud registration method and system based on normal vector constraint correction | |
| Zhu et al. | A calculation method of phenotypic traits based on three-dimensional reconstruction of tomato canopy | |
| CN115187803A (en) | Positioning method for picking process of tender shoots of famous tea | |
| CN116523898A (en) | Tobacco phenotype character extraction method based on three-dimensional point cloud | |
| Zhu et al. | A method for detecting tomato canopies’ phenotypic traits based on improved skeleton extraction algorithm | |
| Gao et al. | Maize seedling information extraction from UAV images based on semi-automatic sample generation and Mask R-CNN model | |
| CN111768413B (en) | Plant three-dimensional point cloud segmentation method and system | |
| Krause et al. | A guided multi-scale categorization of plant species in natural images | |
| CN108108700A (en) | A kind of characteristic area recognition methods of the pig based on peg conversion | |
| Sodhi et al. | Robust plant phenotyping via model-based optimization | |
| CN114494586B (en) | Lattice projection deep learning network broadleaf branch and leaf separation and skeleton reconstruction method | |
| CN116805413A (en) | Automatic calculation method for oil tea seedling stage phenotype characteristics based on three-dimensional point cloud | |
| CN117132802A (en) | Method, device and storage medium for identifying field wheat diseases and insect pests | |
| CN114782455A (en) | Cotton row center line image extraction method for agricultural machine embedded equipment |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |